Is it the case that all GCM computer models projected a steady increase in temperature for the period 1990-2008, whereas in fact there were only 8 years of warming were(sic) followed by 10 years of stasis and cooling.
Correct answer - Yes
If so, why is it assumed that long-term climate projections by the same models are suitable as a basis for public policy making?
Correct answer - Bad advice
It is not the case that all GCM computer models projected a steady increase in temperature for the period 1990-2008.
As noted above, air temperatures are affected by natural variability. Global Climate Models show this variability but are not able to predict when such variations will happen.
The Global Climate Model data presented in the IPCC Fourth Assessment are averages of many individual simulations. By averaging across simulations natural variability is 'smoothed over' and the result shows only the underlying trend due to large-scale forcings such as greenhouse gases. This is illustrated below. The coloured lines are individual 'realisations' or simulations of global average temperature over the period 1950 to 2020 using a particular model (called 20C3M). The dark line is the average of the individual realisations.
The figure below shows that GCM simulations do capture the decadal patterns of variability evident in the temperature record. They do not predict a steady, uninterrupted increase in air temperatures. The left panel shows two periods - 1977-1985 and 1981-1989 - in the global average air temperature record where no substantial warming was observed, although they are embedded in the longer term trend that does show substantial warming. GCMs reflect this type of pattern. The right panel shows a GCM-based projection of 21st century global average air temperature using a single realisation. Note that the 2001-2010 period and the 2016-2031 period show no significant trend although the century-scale trend is one of strong warming - between 3 and 4°C.
Globally averaged surface air temperature for land and ocean based on the data set by Smith et al 
One realisation of the globally averaged surface air temperature from the ECHAM5 coupled climate model forced with the SRES A2 greenhouse gas increase scenario for the 21st century
Therefore, GCMs can and do simulate decade-long periods of no warming, or even slight cooling, embedded in longer-term warming trends.
A simple question. Model forecasts are supposed to become more accurate the further into the future from the initiation of the model we get. How can that be true for homogenised wrong model projections that increasingly diverge as is obvious even in this illustration that only shows to 2020?
If all the model projections were aligned at a common point such as the temperature in June 1988, how diverse would the results be at 2020? That there is no common point suggests graph-craft, manipulation to show the best result.
Models are useless for climate prediction because they don't predict.
Models are useless for projection because assumptions are obligatory for climate factors such as atmospheric nuclear bomb detonations, solar variation, ocean oscillations, air current changes, aerosol variation, cloudiness, water vapour, precipitation, ice extent, volcanic and tectonic activity, land use change and others.
Aspects of climate science are scorned because the science has become a political football. It gets funding inappropriate to its value. Opinion and modelled projections take preference over reality. Climate science is advancing. Maybe in a few decades it will be somewhere near useful for long term policy decision making. As of now, it isn't. Bigger computers merely get the wrong results faster.
In the situation we have at present, insufficient data to compute, ongoing investment in climate projection is akin to investment in astrology.
CO2 is a component of the atmosphere that as a whole moderates climate. Warmed air rises and cools. That is the biggest climate temperature factor. Yet a gas of 0.038% of the whole atmosphere that reacts to radiation that is barely identifiable as measurable heat is given higher importance. Human additions at 3.4% of ~2 parts per million p.a. additions to 0.038% is fancied to influence the climate detrimentally. 0.07% is fancied to cause the sky to fall.
Finally, if the aforementioned is insufficient to show the wrongness of the IPCC support for CO2 "greenhouse" hypothesis, then the following should.
Falsification of the Atmospheric CO2 Greenhouse Effects Link
A non technical summary that should be read in conjunction with the full paper - link. The average Joe/Jolene may be overwhelmed by the mathematics. Skip them. They are the mathematically logical steps to illustrate and evidence the accompanying texts. The document is not difficult to understand and the points covered leave little room for dispute.
From the n-t summary:
Section 3.6 the classic hypotheses of Fourier, Tyndall and Arrhenius are analysed in detail, followed by modern versions of it, and it is concluded that :
• In the 70s, computer simulations of the "global climate" predicted for a doubling of the CO2 concentration a temperature rise of about 0.7 – 9.6 degrees Kelvin.
• Later computer simulations pointed towards a null effect.
• In the IPCC 1992 report, computer simulations of the “global climate" predicted a global temperature rise of about 0.27 - 0.82K per decade.
• In the IPCC 1995 report, computer simulations of the “global climate" predicted a global temperature rise of about 0.08 - 0.33K per decade
• In 2005, computer simulations of the “global climate" predicted for a doubling of the CO2 concentration a global temperature rise of about 2 - 12K, whereby six so-called scenarios have been omitted that yield a global cooling.
To derive climate catastrophes from these computer games and to scare mankind to death is a crime.
Section 5 is the final section of the paper and contains the ‘Physicist’s Summary’, which the reader of this non-technical summary is again urged to review in its entirety. Simply quoting these few lines do an injustice to the entire paper, but set the tone for discrediting the fallacy the UN IPCC is perpetuating, aided in no small measure by many a skeptical scientist who also fails to grasp the fallacy of the so-called greenhouse effect with its double-counting of radiant energy.
Ends with: “The natural greenhouse effect is a myth, not a physical reality. The CO2-greenhouse effect, however, is a manufactured mirage.
Horrific visions of a rising sea level, melting pole caps and spreading deserts in North America and Europe are fictitious consequences of a fictitious physical mechanism which cannot be seen even in computer climate models.
More and more, the main tactic of CO2-greenhouse gas defenders seems to be to hide behind a mountain of pseudo-explanations that are unrelated to an academic education or even to physics training.
The points discussed here were to answer whether the supposed atmospheric effect in question has a physical basis. It does not.
In summary, no atmospheric greenhouse effect, nor in particular a CO2-greenhouse effect, is permissible in theoretical physics and engineering thermodynamics.
It is therefore illegitimate to use this fictitious phenomenon to extrapolate predictions as consulting solutions for economics and intergovernmental policy.”
See also: jennifermarohasy.com/blog/2009/03/radica.